Abstract
Human mitochondrial DNA (mtDNA) data from 18 populations have been carefully reexamined. A phylogeny of 77 mtDNA types found among the 1389 individuals analyzed for restriction fragment length polymorphisms (RFLPs) was established using the parsimony principle and compared to a UPGMA tree of the 18 populations. Both analyses agreed in separating African samples from the other populations, though the mtDNA type phylogeny suggested close relations between Africans and other continental groups. Conformity of observed mtDNA type frequency distributions with the “infinite allele” model was studied for 31 human populations. Several Oriental and Caucasoid populations were found to be overly homogeneous, generally due to an elevated frequency of one particular type. Contrastingly, all African samples conformed to the neutral model of populations at equilibrium and presented more diversified distributions. This suggested that part of the apparent African divergence was due to heterogeneous evolutionary processes and confirmed that some diversity reducing factors were at work in Caucasoids and Orientals. Several nonexclusive hypotheses accounting for the rejection of the neutrality tests were discussed. Alternative hypotheses concerning modern human emergence were also reviewed in the light of present results.